Earth Observation in Mining Activities: Phosphate Deposits in Morocco

Identifying surface minerals remotely is crucial for natural resource management. Remote sensing allows for efficient mapping of mineral distributions, which is beneficial for the mining and energy sectors, reducing the costs and difficulties of on-the-ground exploration. Moreover, mineral dust from sources like mines and deserts significantly impacts climate, air quality, and ecosystems. Hyperspectral sensors, such as EMIT, are often used to identify dust sources and track their effects on atmospheric conditions, weather, and human health. However, challenges arise when certain minerals are nearly featureless in the sensor's designed wavelength range, or when spectral features overlap, making it difficult to distinguish between minerals. To address the limitations of sensors operating in visible and infrared ranges, this study adopts a novel approach. The ECOSTRESS sensor, originally developed to monitor Earth's surface temperature, is repurposed to explore its potential for mineral detection and characterization. This method aids in identifying potential indicators of rare earth elements, which are critical for various technological applications. In this study, the NDVI was used to characterize the vegetation patterns and types in the Youssouffia region over five months in 2023. ECOSTRESS and EMIT images were acquired after selecting the appropriate preprocessing level. For ECOSTRESS, swath data was converted to GeoTIFF using a Python script, followed by band stacking, clipping, and conversion from emissivity to reflectance. In the hyperspectral analysis, mineral spectral features were extracted and compared against spectral libraries, with the results from both sensors compared. While the use of EMIT in the hyperspectral analysis of Youssouffia shows promise, it has significant limitations when studying minerals with overlapping spectral features or those that are featureless in certain wavelengths (e.g., SWIR). To overcome these challenges, ECOSTRESS data was used for spectral analysis, yielding valuable results in distinguishing minerals like quartz and apatite. Although both EMIT and ECOSTRESS have technical shortcomings in hyperspectral and spectral analysis, reexamining the adopted methodology with other sensors and satellites could increase the reliability of this study. Ground truth verification and field studies would enhance the robustness of the results and provide a foundation for further applications, such as using these findings for AI training.